Sebastián Valenzuela

How much do experts on the information environment agree about the scale, spread, and effects of misinformation? We drew on three annual waves of the International Panel on the Information Environment expert survey to examine agreement with seven popular claims about misinformation (2023: N = 262; 2024: N = 413; 2025: N = 442; spanning 76 countries). Agreement varied widely across statements and was stable across waves. About 60% of experts agreed that falsehoods spread faster than the truth and that misinformation is widespread. About half agreed that misinformation has a substantial influence on behavior and that people readily believe what they see online; about 40% agreed that more people have become misinformed over time; and only 25% agreed that misinformation is primarily a social media problem. The surveys also revealed a persistent divide between experts who saw misinformation as widespread and consequential and those more skeptical, with sharpest disagreement on its behavioral impact and increase over time. In 2025, some experts distinguished between misperception prevalence and misinformation severity. Overall agreement varied by respondent background: agreement was higher among experts on developing countries than those on developed countries, higher among STEM researchers than social scientists, and lower among political scientists than communication scholars. Expert consensus on misinformation is partial, varies across claims and experts, and is shaped by geography and discipline.

Generative AI produces information on demand, raising the question of why one would ever learn something that a chatbot can always retrieve. In this paper, we introduce Belief in AI Sufficiency (BAS): the belief that acquiring and retaining knowledge is less valuable because AI can provide it when needed. BAS represents a qualitatively distinct conceptual domain from the Google effect, cognitive offloading, and the “News Finds Me” perception. Using four preregistered studies in the United States (development, N = 307; validation, N = 903; experiment, N = 767; test-retest, N = 298), we present evidence for a 12-item BAS measure with a strong general factor structure. In the validation survey, higher BAS was associated with greater reliance on heuristic and less systematic processing, and with reduced factual knowledge. In the experiment, BAS strongly predicted the likelihood to delegate a learning task to AI. The relationship between BAS and the gap between self-reported and measured knowledge was smaller, reflecting reduced accuracy rather than an illusion of knowledge. Finally, BAS scores were approximately equal in mean level and showed moderate rank-order stability across a 3-month period. BAS is one way to measure how far AI is becoming central to the study of how people learn.

Digital journalism scholars have extensively discussed the implications of artificial intelligence (AI) and, more recently, large language models (LLMs) on journalists’ core practices (Arguedas & S...

The information environment is under great pressure, and so are the researchers who study it. People increasingly receive information already sorted for them, by recommendation algorithms or by generative AI systems that answer directly rather than link to sources. How do experts around the world assess threats to the information environment? In May and June 2026, the IPIE surveyed 470 researchers with expertise on 71 countries in the fourth and most diverse wave of its annual expert survey. Nearly half the experts study developing economies. This Summary for Policymakers presents the main findings of Trends in the Information Environment: 2026 Expert Survey Results (SR2026.3). The report reaches four main conclusions: AI summaries deliver speed at the expense of visits to news websites and viewpoint diversity. Conditions for research on the information environment have worsened by every measure, most notably in the USA. Experts are most worried about unaccountable platform power, followed by misinformation and polarization. Pessimism stands at a record level, and it is deepest where research conditions are rated better. The full report presents these findings in detail, with further observations on the technologies experts view most favorably and on how perceived risks from AI-generated content compare with experimental evidence.

In May and June 2026, the IPIE’s fourth annual expert survey included 470 experts studying the information environment across 71 countries, almost half of whom focused on developing economies. Pessimism has climbed to its highest level yet. This year, 76% of experts expect the information environment in their country of expertise to worsen over the coming year, up from 53% in 2023. For four years running, experts agree on what a healthy information environment needs most: accurate information and a diversity of voices. Most experts, 76%, say people can find diverse viewpoints online, but only 28% think a majority of people in their country trust the main news outlets to report politics accurately. Lack of platform accountability and misinformation are seen as the biggest threats to the information environment, and experts consider generative AI a less urgent threat. For three years in a row, experts expect generative AI to make the information environment worse over the next five years, with AI-generated video their biggest worry and AI-generated text their smallest. Experts expect AI summaries in search engines to speed up how quickly people find what they are looking for, but to cut traffic to news sites and narrow the range of viewpoints they see. Search engines again rank as the technology seen most positively, 80% positive, while views on social media, generative AI, and recommender systems are mixed to negative, 35–45% positive. Research funding remains the most frequent barrier to studying the information environment, 67%, ahead of data access, 60%, and roughly a third of experts report government-related pressures on their work. Behind these overall figures, experts in developing economies and the USA show similar patterns: they rate threats higher, report more government censorship and surveillance, and view the effects of technologies, including AI, similarly. USA experts are nearly twice as likely as experts in all other countries to report government control of funding and pressure to align research with a political agenda as main barriers to studying the information environment. Experts fall into three different groups based on their concerns: about half worry about almost every threat to the information environment, a third worry mainly about platform accountability and misinformation but not AI, and one in ten hardly worry at all.

Digital literacy is often proposed as a remedy for misinformation, yet empirical evidence regarding its effectiveness remains mixed. One reason for these inconsistencies lies in the multidimensional nature of both literacy and misinformation. This article focuses on algorithmic literacy, a distinct and increasingly central form of digital literacy in platform-mediated communication, and examines how its cognitive and behavioral dimensions relate to multiple facets of misinformation engagement. Drawing on two probabilistic survey studies conducted in Chile—a nationally representative face-to-face survey ( N = 1,200) and a random-digit-dial telephone survey of urban adults ( N = 2,265)—we analyze how cognitive (algorithmic awareness and knowledge) and behavioral (strategic actions to navigate algorithmic systems) dimensions of algorithmic literacy relate to perceived misinformation exposure, misinformation sharing, objective and subjective news knowledge, and knowledge miscalibration. Across both studies, a consistent pattern emerges: algorithmic literacy is more strongly associated with cognitive outcomes—especially perceived misinformation exposure and knowledge—than with behavioral outcomes like sharing. Higher algorithmic awareness consistently predicts greater perceived exposure, suggesting heightened recognition, while its link to sharing behavior is less consistent. These studies highlight the relevance of distinguishing cognitive from behavioral dimensions of literacy across different stages of misinformation engagement.

Political information environments around the world are facing the twin challenges of misinformation and disinformation, which degrade the quality of public discourse, undermine political stability, and weaken social cohesion. It is therefore important to understand the conditions under which countries are at risk from misinformation and disinformation. In recent years, several indexes have been developed by scholars, think tanks, advocacy groups, and technology companies aimed at diagnosing and comparing these risk and protective conditions. Yet, systematic cross-index comparison remains limited. This Synthesis Report (TP2026.1) addresses this gap by developing a typology of existing measures, providing a systematic comparative analysis of ten published indexes, and offering a fitness-for-purpose guide for researchers, policymakers, and index producers. Our analysis reveals three key insights: The concepts of vulnerability and resilience provide a foundational framework for understanding how political misinformation and disinformation are amplified or mitigated and for comparing countries over time. However, because most indexes rely on structural, perceptual, or behavioral proxies rather than direct measures, their scores should therefore be read as comparative diagnostics rather than estimates of disinformation prevalence. Several indexes offer valuable and complementary insights, but they diverge in their theoretical assumptions, methodologies, and the aspects of the political information environment they emphasize. Caution is therefore required to avoid overgeneralizing cross-national rankings, and triangulation across complementary measures is often advisable. Although the ten indexes considered in this report were designed to address political misinformation and disinformation, their underlying logic can be applied to the broader task of measuring political information environments and information integrity. By evaluating the strengths, good practices, and shortcomings of these ten indexes, this report highlights the critical dimensions and measures that can guide a theoretically informed, methodologically robust, and globally sensitive next-generation index. It concludes with guidance for using index results responsibly.

This report synthesizes experimental evidence on the effects of generative artificial intelligence (GenAI) misinformation on individuals’ ability to evaluate information and the effectiveness of interventions designed to mitigate its influence. It draws on a meta-analysis of experimental studies (60 effects from 24 publications and 33,801 participants) published between 2018 and 2025. The findings indicate a divergence in public responses to different modalities of GenAI-produced misinformation (referred to in the report as GenAI misinformation), alongside substantial heterogeneity across studies. More recent research tends to report higher perceived accuracy and credibility of textual GenAI misinformation, whereas individuals’ perceptions of visual GenAI misinformation—primarily deepfakes—have become more skeptical over time. These patterns are associated with the time of data collection rather than model capacity and are accompanied by wide variation in effect sizes across contexts. The report also evaluates two countermeasures: preventive corrective information and content labeling. In studies conducted after 2020, corrective information yields small-to-moderate (0.17 to 0.43 standard deviation units) reductions in the perceived accuracy or credibility of GenAI misinformation. Pooled estimates in this subset are significant and comparatively stable, though prediction intervals show substantial variation, especially when outcomes focus on detection rather than evaluative judgment. Content labeling interventions vary more than corrective information. Althoughlabels are associated with modest average reductions in perceived credibility, effect sizes vary widely across modalities, designs, presentations, and contexts,and prediction intervals often include null effects. Key conclusions from the synthesis are as follows: Textual GenAI misinformation may currently pose greater persuasive risks than visual GenAI misinformation. Preventive corrective information appears to be more a consistently effective countermeasure. Content labeling policies yield highly variable outcomes and require careful design and evaluation. The available evidence base is geographically narrow and methodologically constrained, with most studies focusing on populations in Western Europe and North America Based on these findings, policy responses should consider the following: Prioritize addressing text-based GenAI misinformation. Treat preventive corrections as a core intervention. Approach content labeling cautiously